Model Router
Model Router — The Right Model per Task.
Different jobs need different models. The router picks a suitable one per task — fast models for short operations, stronger models for planning and analysis — so you do not have to think about model names to get good results.
Straight answers
- What is a model router?
- It is the component that decides which AI model handles a given task, based on the type of work, the required depth and your configuration.
- Can I override it?
- Yes. You can pin a preferred model for a task, agent or workspace.
What you get
200+ models
A broad catalogue available inside the app.
Task-aware
Routing based on the kind of work requested.
Manual pinning
Force a specific model when you need consistency.
Per-agent settings
Each agent can prefer its own model.
Fallbacks
Work continues if a provider is unavailable.
Cost awareness
Cheap work stays on cheap models.
Why routing matters more than any single model
Sending every request to the largest model is slow and wasteful; sending everything to the smallest one produces poor plans. Routing gets the balance right automatically across a long working day.
Frequently asked
- Does it lock me to one provider?
- No — model choice is configurable and multiple providers are supported.